Detection of coronary artery disease using an electronic stethoscope

Abstract

The disclosure describes an electronic stethoscope system that automatically detects coronary artery disease in patients. The system uses an electronic stethoscope to record acoustic data from the fourth left intercostal space of a patient. A processing technique is then applied in order to filter the data and produce Fast Fourier Transform (FFT) data of magnitude versus frequency. If a bell curve is identified in the data between a predefined frequency range (e.g., 50 and 80 Hz) with a peak magnitude of greater than a predefined threshold (e.g., 2.5 units), the system automatically provides an output indicating that the patient is likely to have 50 to 99 percent stenosis of the coronary artery. If no bell curve is present, the patient may have artery stenosis of less than 50 percent. An interface module may be used to transfer diagnosis information to the stethoscope and data to a general purpose computer. This inexpensive and quick system may improve the chances for early detection and patient survival of coronary artery disease.

Description

This application claims the benefit of U.S. Provisional Application No. 60/670,905, filed Apr. 13, 2005, the entire content of which is incorporated herein by reference.

TECHNICAL FIELD

The invention relates to medical diagnostic devices and, more particularly, electronic stethoscopes.

BACKGROUND

Coronary artery disease affects the lives of millions of people, and may affect the health of a patient without warning. Detection of coronary artery stenosis involves patient history, physical examination, stress testing and possibly a coronary angiogram. Beyond history and physical examination, the diagnostic technique is associated with significant cost and risk. Although the stress test is the most frequently ordered test to detect possible coronary artery disease, sensitivity and specificity of the stress test vary greatly from 40 percent to 90 percent, depending upon whether there is single or multi-vessel disease.

During routine physical examination in a clinic office, physicians and other medical providers use a stethoscope. It is inexpensive, easily portable, relatively comfortable, and safe. Advancements in digital technology have led to the production of electronic stethoscopes that can amplify sound, record patient data and transmit data to a computer for further processing. The transmitted data can be used to plot a phonocardiogram, improve patient records and even perform automated heart sound diagnosis.

Within the electronic stethoscope, an acoustic sensing device is employed to transmit the raw sound data from the patient to electronics within the stethoscope. The raw data from these sensors contains a plethora of acoustic information emanating from the thorax that includes not only heart valve and lung sounds but also acoustic information within the stethoscope sampling frequency and signal to noise ratio. Once in digital form, the data from the stethoscope sensor is filtered so that it sounds like a mechanical stethoscope.

SUMMARY

The invention is directed to an electronic stethoscope system that detects coronary artery disease in patients. The system uses an electronic stethoscope to record acoustic data from the chest. A processing technique is applied to the data in order to produce Fast Fourier Transform (FFT) data of magnitude versus frequency.

If a bell curve is identified in the data within a predefined frequency range with a peak magnitude of greater than a predefined threshold, the system produces an output to indicate that the patient is likely to have 50 to 99 percent blockage, or stenosis, in the coronary artery. As described herein, it has been identified that a blockage exceeding 50 percent manifests in a bell curve in the range of 50-80 Hz within the FFT data. As a result, in one embodiment, the system automatically determines that the patient is likely have 50 to 99 percent blockage, or stenosis, in the coronary artery when a bell curve is identified on the graph between 50 and 80 Hz with a peak magnitude of greater than 2.5 units. If no bell curve is present, the patient may have stenosis of less than 50 percent in the coronary artery.

Since artery stenosis can be left undetected until blood flow is severely restricted within the artery, a quick and non-invasive technique described herein would greatly increase the number of diseased patients that could be identified before their health deteriorates. Implementing the described processing methods into an electronic stethoscope may allow early warning for patients without disease symptoms. Positively identified patients could then participate in more invasive techniques to specifically identify the extent of the disease. This technique may also assist in preventing normal patients from having unnecessary expensive and dangerous techniques, i.e. angiography, performed.

In one embodiment, the invention provides method for detecting coronary artery disease using a stethoscope comprising placing an acoustic sensor over the fourth left intercostal space of a patient's chest, recording acoustic data from the acoustic sensor, applying one or more filters to the acoustic data and calculating a Fast Fourier Transform (FFT) of the data to produce FFT data, and automatically analyzing the FFT data to identify a bell curve within a predefined frequency range indicative of coronary artery disease.

In another embodiment, the invention provides an electronic stethoscope comprising an acoustic sensor, a memory that stores acoustic data from the acoustic sensor and instructions for processing the acoustic data, and a processor that processes the acoustic data to identify a bell curve in FFT data between a predefined frequency range indicative of coronary artery disease.

Although the invention may be especially applicable to detecting left anterior descending coronary artery stenosis, the invention alternatively may be applied to blockages in other arteries or the phenomenon of restinosis after a stent has been implanted to open the artery.

In various embodiments, the invention may provide one or more advantages. For example, the use of an electronic stethoscope to diagnose coronary artery disease may allow any patient to be screened due to the widespread use of electronic stethoscopes and the little amount of time this technique would require. During a routine physical examination, a physician may decide to perform this technique with only the equipment on his person and a few minutes of time. The physician may decide to have the system automatically produce a diagnosis based upon the processed data, or the physician may choose to print the data from a local computer to analyze the FFT graphed data himself.

This technique to identify coronary artery disease may also benefit patients who would normally be subjected to inaccurate stress testing or invasive angiography. Some patients with less than 50 percent artery occlusion, who would have been subjected to one of these tests, would be cleared of any addition testing. This may save patients from possible pain, injury, procedural costs, and unnecessary time. Strained health care facilities may also save personnel time and procedural costs.

In addition, the patient may modify the parameters of the processing technique to accommodate a variety of patients. An external module may be used to communicate with the electronic stethoscope to modify settings or upload updated processing tools. The external module may also be used to download data from the electronic stethoscope, transfer data to a general purpose computer, and recharge the battery of the electronic stethoscope.

The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram illustrating an electronic stethoscope system placed on the chest of a patient to diagnose coronary artery disease.

FIG. 2 is functional block diagram illustrating the method to record and process acoustic data from the stethoscope.

FIG. 3 is functional block diagram illustrating the method to diagnose coronary artery stenosis from the processed acoustic data.

FIG. 4 shows data charts from two separate normal patient data sets after processing is completed.

FIG. 5 shows data charts from a patient with 90-99% artery stenosis. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data.

FIG. 6 shows data charts from a patient with 50-75% artery stenosis. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data.

FIG. 7 shows data charts from a patient with 75-90% and 100% artery stenosis in different artery sections. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data.

FIG. 8 shows data charts from a patient with 50-75% and 100% artery stenosis in different artery sections. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data.

DETAILED DESCRIPTION

FIG. 1 is a schematic diagram illustrating an electronic stethoscope system 10 to diagnose coronary artery disease. In the example of FIG. 1, system 10 includes an electronic stethoscope 14 applied to the chest of patient 12. Electronic stethoscope 14 may communicate with interface module 16, which may be internal or external to stethoscope 14, in order to transfer data from stethoscope 14 to a general purpose computer (either wirelessly or via conventional wired communication) or to modify diagnosis parameters of stethoscope 14.

With reference to FIG. 1, electronic stethoscope 14 includes an acoustic sensor, earphones, and an electronic processing element. The acoustic sensor is located at the distal, or patient, end of stethoscope 14. This sensor turns sound waves created within patient 12 into electrical signals that may be used by the stethoscope. Various types of sensors that may be used include microphones or piezoelectric crystals, or any sensor capable of detecting audible or inaudible vibrations associated with hemodynamics or arterial elasticity in the coronary arteries. In some embodiments, a plurality of acoustic sensors may be implemented to acquire patient data. The acoustic data may be filtered or amplified before reaching the earphones located at the proximal, or physician, end of electronic stethoscope 14. The physician may be able to modify the sound by adjusting the volume or certain frequencies.

The acoustic sensor of stethoscope 14 may be placed on the chest of patient 12. In particular, the acoustic sensor may be placed over the fourth left intercostal space. As blood flows through arteries, blockages within an artery lumen may cause the blood flow profile to change from laminar flow to turbulent flow as the blood passes faster over these obstructions. This turbulent flow produces sounds capable of being detected by the sensor within stethoscope 14. From this fourth left intercostals space, the sounds most applicable to coronary artery stenosis are audible from the chest. In some embodiments, other locations may be more appropriate when attempting to record acoustic data from this or other arteries.

The electronic processing element of electronic stethoscope 14 controls the function of the stethoscope. Acoustic data covering systole and diastole is acquired from the acoustic sensor, at which time the analog signal may be amplified or modified for listening by the physician. The signal is sent in real time to the earphones of the stethoscope for patient diagnosis purposes. While the signal is sent to the earphones, it may also be recorded within the processing element. Acoustic data may be stored in the memory for off-line analysis.

In this embodiment, the acoustic data recorded by stethoscope 14 may be downloaded by interface module 16. Communication between stethoscope 14 and interface module 16 may be accomplished through a Universal Serial Bus (USB) connection, IEEE 1394 connection, wireless telemetry connection, or some other transfer connection. Interface module 16 may then transfer the data to a general purpose computer, i.e. a desktop, notebook, or hand-held computer. The data may then be processed using computational software, i.e. MATLAB, to generate data for diagnosis. This data may be viewed on a screen, transferred to another computing device, or printed on paper. The physician may then make a diagnosis of the patient.

In some embodiments, acoustic data may be processed within the processing element. When desired by the physician, stethoscope 14 may automatically process the acoustic data and make a recommended diagnosis of the patient. By following the diagnosis technique to analyze the processed data, stethoscope 14 may present the diagnosis information through an LED indicator, LCD screen, or audio message to the physician. In this manner, the physician may be able to quickly analyze the acoustic data without the use of multiple devices. In addition to the automated diagnosis, the physician may have the option of manually analyzing the data as well.

In other embodiments, acoustic data may be automatically processed by sending acoustic data from stethoscope 14 to interface module 16. Data processing and analyzing may occur at this module or a host computer to conserve battery life or improve processing speed. The recommended diagnosis may then be delivered to the physician through a user interface on module 16 or the host computer.

Some embodiments of electronic stethoscope 14 may be able to monitor a variety of patient information and suggest a coronary artery analysis based upon certain indicators. In this case, stethoscope 14 may be able to assist the physician in detecting artery blockages or in-stent restenosis in patients with limited signs or symptoms.

FIG. 2 is a functional block diagram illustrating the method to record and process acoustic data from stethoscope 14. The technique begins when the physician turns on electronic stethoscope 14. Once initiated, the acoustic sensor of stethoscope 14 is placed over the fourth left intercostal space of the chest (18). Once positioned, stethoscope 14 begins recording acoustic sounds produced over systole and diastole within the chest (20). Stethoscope 14 begins processing the signal by applying an analog to digital converter to produce digital data representative of an acoustic signal (22). Next, stethoscope 14 downsamples the digital data by one half to create a manageable data set (24). Stethoscope 14 also applies a first filter (“Filter 1”) to the data in order to remove non-stenosis frequencies from the data (26), after which the stethoscope applies a low-pass filter (“Filter 2”) to the data to eliminate frequencies above 100 Hz (28). Then, stethoscope 14 calculates a Fast Fourier Transform (FFT) of the data (30), in which the resultant FFT data may analyzed and/or graphically plotted with frequency on the x-axis and magnitude on the y-axis (32). Fourier analysis is based on the concept that signals can be approximated by a sum of sinusoids, each at a different frequency. Plotting the harmonic magnitudes on the y-axis (no units) and the frequency of the harmonic on the x-axis (Hz) generates a frequency spectrum. The spectrum is represented as a set of vertical lines or bars. The frequency has units of Hz and the magnitude has non-dimensional units. It can therefore also be referred to as a coefficient. Stethoscope 14 may then calculates a maximum and minimum bar from the FFT plot (34) and produces diagnostic data for the coronary artery (36).

For exemplary purposes, this functional block diagram shows the process stethoscope 14 would use to automatically process acoustic data for the diagnosis of coronary artery disease. In some embodiments, this process may be performed automatically by interface module 16 or manually by a clinician or technician with the aid of a personal computer.

FIG. 3 is a functional block diagram illustrating an exemplary method in which stethoscope 14 automatically diagnoses coronary artery stenosis from the processed acoustic data.

The diagnostic data containing the coronary artery data (36) needs to be analyzed for indications of stenosis. This may be done manually by a clinician or automatically by stethoscope 14, as illustrated by this example. First, stethoscope 14 determines whether a bell shaped curve exists within a particular spectral region of the FFT data (38). In particular, the stethoscope 14 determines whether the onset of the curve (i.e., the lower frequency where the upslope of the curve rises) occurs substantially at or closely after 50 Hz (40). If this is true, stethoscope determines whether the downslope of the bell shaped curve ends substantially at or before 80 Hz, i.e., that the bell shaped curve is bounded by this spectral region (42). If so, stethoscope 14 determines whether the FFT peak of the bell curve bounded by this region exceeds than 2.5 units (44). If all four of these criteria are met, stethoscope 14 (or interface module 16) outputs an indicator, e.g., a message, that the Left Anterior Descending (LAD) portion of the coronary artery is likely to have more 50 to 99 percent stenosis preventing blood flow and that coronary disease may be present. (46). If one or more of the presented criteria is not satisfied, then the stethoscope outputs an indicator, e.g., message, that the patient is likely to have less than 50 percent of the LAD portion of the coronary artery blocked (48).

For exemplary purposes, this functional block diagram shows the process stethoscope 14 would use to automatically diagnose coronary artery disease from processed acoustic data. In some embodiments, this process may be performed automatically by interface module 16 or manually by a clinician or technician with the aid of a personal computer. In the cases of manual diagnosis, the FFT data may be graphed or plotted for the clinician to observe the data as a whole, and the identified bell curve may be highlighted or colored differently from the other portions of the graph. For example, a bell shaped group of data would be identifiable between 50 and 80 Hz on the graph in the case of a patient with greater than 50 percent artery stenosis.

In some embodiments, the criteria described in FIG. 3 may be determined in different order. For example, after the bell shaped curve is identified in block 38, the next criteria of the curve identified may be a FFT peak of greater than 2.5 units as shown in block 44. While each criterion is assessed to determine which indicator to output relative to the likely presence of LAD stenosis, the order in which they are assessed is not as critical.

In other embodiments, certain criteria may indicate a particular percentage of stenosis and enable the diagnosis to separate a patient into more than two diseased states. A more sensitive diagnosis indicating more percentage levels of stenosis may allow for certain patients to wait before undergoing treatment while patients diagnosed with a greater percentage of stenosis would be urged to have immediate surgical intervention.

In other embodiments, LAD coronary artery stenosis may be detected through a slightly different method. Calculating the sum of the energy under the bell curve described above may allow the stethoscope to present an interval of probability of 50 to 99 percent stenosis. Calculating the energy from the data may enable more accurate diagnoses. Further, the summed energy information may allow the stethoscope to provide increased separation of the degree of stenosis. For example, the patient may be identified as likely having 25 to 50 percent, 50 to 75 percent, or 75 to 99 percent stenosis of the LAD or alternate portion of the coronary artery. This information may be beneficial to both the patient and physician.

In addition, the specific numerical indicators for detecting stenosis in the LAD may vary depending on the condition of the patient. For example, the FFT peak in block 44 may be greater than 2.5 units to indicate stenosis over 50 percent patients over 60 years of age, whereas the FFT peak in block 44 may be greater than 3 units to indicate stenosis over 50 percent in patients younger than 60 years of age. Factors such as age, gender, height, weight or history may be used in determining the exact criteria numbers. In general, the exact coefficients used in these criteria may be modified to best match the patient.

This technique may be particularly useful for initial screening of patients for LAD stenosis indicating coronary artery disease during a routine exam with a primary care physician. Patients failing to meet every criterion would not need further tests until at some time they show stenosis greater than 50 percent. Patients diagnosed with LAD stenosis greater than 50 percent may be directed toward further testing, such as an angiogram or angioplasty.

Various embodiments of the described invention may include processors that are realized by microprocessors, Application-Specific Integrated Circuits (ASIC), Field-Programmable Gate Arrays (FPGA), or other equivalent integrated logic circuitry. The processor may also utilize several different types of storage methods to hold computer-readable instructions for the device operation and data storage. These memory and storage media types may include a type of hard disk, random access memory (RAM), or flash memory, e.g. CompactFlash or SmartMedia. Each storage option may be chosen depending on the embodiment of the invention.

EXAMPLES

FIGS. 4 through 8 illustrate some of the results of the disclosed technique for diagnosing coronary artery stenosis when performed on patients that were already scheduled for a routine coronary angiogram due to typical symptoms or a positive stress test. Patients were excluded from this study if they had prior coronary artery bypass or cardiac transplant surgery, persistent atrial fibrillation, or any type of cardiac device including pacemakers and prosthetic valves.

To objectively quantify stenosis, angiographic data was collected from the final angiographic report. Coronary artery stenosis was classified as 25-50 percent, 50-75 percent, 75-90 percent, and 90-99 percent occlusion pre-intervention at the referral center. Data from 55 random patients was collected (age range 29-84, mean 57; 30 female, 25 male, Body Mass Index range 21-50, mean 30). Seventeen patients had normal angiograms, 6 had stenosis in coronary arteries other than the LAD, 11 had 25-50 percent stenosis, 5 had 50-75 percent stenosis, 5 had 75-90 percent stenosis, and 7 had 90-99 percent stenosis. A 41 year-old patient tested positive for pregnancy, another patient was diagnosed with MoyaMoya disease (had diffuse coronary spasm), a third patient refused the angiogram after being enrolled, and a fourth patient had a stent previously placed. All four patients were removed from our study.

A Littmann Model 4000 electronic stethoscope with recording frequencies within the human range of hearing (20-4000 Hz) was used to collect the data. This stethoscope can record a maximum of eight seconds of acoustic data per recording. Eight second recordings were collected at the 4^th left intercostals space of a supine patient. A pre-angiogram was first taken and if the patient had an angioplasty, the same stethoscope was used to collect data post-angioplasty.

The 4 lics thorax sounds were sampled as a series of voltages representing the sound amplitude of the acoustic signature. A filter was applied to the data to remove frequency information not related to stenosis in the LAD. A lowpass filter was then applied to the signal to eliminate noise and narrow the bandwidth to less than 100 Hz. The techniques included an analysis of acoustic information within the 20-100 Hz range of both the systolic and diastolic segments of the cardiac cycle. The frequency range and use of the entire cardiac cycle was used.

Using MATLAB 6.5, Release 13 (a program used from numeric computation and visualization) a Fast Fourier Transform (FFT) was calculated on the filtered data. A max-min bar was calculated for the resulting FFT data. The max-min bar was chosen so that variability between coefficient magnitudes and frequencies could be examined. Finally, the x-axis was plotted on a logarithmic scale for the 20-100 Hz frequency range.

Interpretation of the graph resulting from the analysis was based on a assessment of four graphical elements: 1) presence of a bell-shaped curve resulting from the variability between coefficient magnitudes and frequency, 2) onset of the bell-shaped curve at 50 Hz, 3) downslope of the bell-shaped curve ends ≦80 Hz, and 4) FFT peak coefficient magnitudes greater than 2.5 units (absolute scale for all analyses) at the maximum of the bell-shaped curve. If all four criteria are met, then the patient was identified to have 50-99 percent stenosis in the LAD. If all four criteria are not met, then the patient was identified as normal.

The graphs of FIGS. 4-8 compares data from real patients with and without coronary artery disease as determined through an angiogram, the current standard in conventional diagnosis. In the case of patients receiving intervention to rectify their disease, data was acquired before the angiogram and after the angioplasty. The same stethoscope was used for each diagnosis of a single patient.

FIG. 4 shows data charts generated for two separate normal patient data sets after processing is completed. The FFT graphical result shows normal coronary arteries between 0 and 100 Hz. The 50 to 80 Hz frequency bands associated with suspected LAD stenosis are highlighted with a rectangular box. As seen in graphs A and B, the coefficient magnitude decays with increasing frequency. In this example, there is no bell-shaped curve, so peak magnitude is irrelevant and the data fails to meet the criteria set forth in FIG. 3.

FIG. 5 shows data charts from a patient confirmed to have 90 to 99 percent artery stenosis in the proximal LAD. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data. In this graph, the bell-shaped curve that may be associated with LAD stenosis was detected as shown. The bell-shaped curve upslope begins at 50 Hz and the downslope ends near 65 Hz. The maximum coefficient magnitude is greater than 2.5 units and is found at 55 Hz, the peak of the bell-shaped curve. The bell-shaped curve disappeared after angioplasty or stenting of the LAD. In this manner, Graph B of the patent taken post angioplasty is similar to a graph of a normal patient.

FIG. 6 shows data charts from a patient confirmed to have 50 to 75 percent stenosis in the first diagonal branch of the LAD. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data. LAD stenosis was identified with the bell-shaped curve in the middle of the 50 to 80 Hz frequency range, and then confirmed via an angiogram. In this case, the maximum coefficient peak was greater than 2.5 units at approximately 65 Hz and decays to 80 Hz. A single drug-eluting stent was placed during angioplasty of this patient. Again, the techniques correctly identified a patient with a diseased coronary artery.

FIG. 7 shows data charts from a patient with 75 to 90 percent stenosis in the LAD and 100 percent artery stenosis in the right coronary artery. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data. The upslope of the bell-shaped curve begins at 50 Hz. The maximum coefficient magnitude is greater than 2.5 units at 55 Hz. The bell-shaped curve ends at 60 Hz. This technique correctly diagnosed a diseased coronary artery in this patient. Graph B shows the absence of the bell-shaped curve after the angioplasty of the LAD was performed, and stenosis similar to that of a normal patient.

FIG. 8 shows data charts from a patient with 50 to 75 percent stenosis in the LAD and 100 percent artery stenosis in the right coronary artery. Graph A shows pre-angiogram data, and graph B shows post-angioplasty data. As shown, the upslope of the bell-shaped curve begins at 50 Hz, peaks above 2.5 units at 65 Hz and downslopes to end before 80 Hz. After angioplasty, graph B shows the disappearance of the bell-shaped curve due to the opening and stenting of the LAD.

Statistical analysis was then performed on the data from the fifty-one patients. If one were to consider the technique overall, it correctly detected both normal and diseased patients in 43 out of 51 cases, or 84 percent of the time.

The data was statistically analyzed by first grouping the data according to whether a patient had any stenosis in the LAD. If a patient had stenosis in other coronary arteries but not the LAD, they were considered normal. The 25-50 percent stenosed patients were included in normal because the angiogram indicated ‘no significant focal stenosis.” There were a total of 34 normal patients. If the LAD was stenosed more than 50 percent, it was called stenosed LAD. There were a total of 17 stenosed LAD patients.

Table 1 shows how the patients were assessed in the angiogram and how well the technique performed:

	TABLE 1
	Normal	Stenosed LAD	Total
	Normal	17	0	17
	Other Artery	6	0	6
	25-50%	9	2	11
	Total Normal	32	2	34
	50-75%	0	5	5
	75-90%	3	2	5
	90-99%	3	4	7
	Total Stenosed LAD	6	11	17

All 17 normal patients were classified as normal—as were all patients with blockage in arteries other than the LAD. In the 25-50 percent stenosis category, 9 out of 11 patients were correctly called normal.

In the Stenosed LAD category, this technique correctly classified 50-99 percent blockage in 11/17 patients. Because of the low sample size within each category, patients within this classification were considered as a whole.

To break it down further, we observed a 94 percent (32/34) success in detecting normal patients. As shown in Table 1, the two patients incorrectly identified as normal were in the 25-50 percent stenosis range. This means that in our data set, the incorrectly classified patients had some stenosis in the LAD. Because we did not use intravascular ultrasound to quantify stenosis, we do not know if the patients were at the lower end (25%) or upper end (50%) of the range.

Patients with LAD (50-99 percent) stenosis were identified 11 times out of 17. This is an observed success rate of 65 percent. Thus, the results indicate the diagnosis technique has 94 percent specificity and 65 percent sensitivity.

This study was conducted to assess the feasibility of this method to detect clinically significant LAD stenosis in patients with symptoms and positive stress test where an interventional procedure was performed. In all cases, the physician who performed the intervention had no knowledge of this study and had the option of using intravascular ultrasound as needed in an individual case. We compared this technique with standard angiography, and found that this technique had a high specificity and negative predictive value for detecting coronary artery disease at the 4 lics in the left anterior descending coronary artery. The high negative predictive value suggests that it may have a role in ruling in clinically significant disease in the LAD coronary artery.

With the use of this technique, 94 percent of the patients without left anterior descending coronary artery disease were identified as having no clinically significant coronary artery disease. These data support the use of this technique a method of ruling in significant disease in the left anterior descending coronary artery. Such information is clinically relevant, since predictive assessment would allow a physician to alert a patient of developing blockage and also allow more time for interventional measures to begin.

Although limited only to the left anterior descending coronary artery, this tool may be useful in screening patients with suspected coronary artery disease, atypical symptoms, family history or multiple risk factors. When the method is applied to raw acoustic information, the resulting analysis can be correlated to stenosis in the left anterior descending coronary artery.

As shown in FIGS. 4 through 8, data from several stethoscope diagnoses may be compared as the condition of a patient develops. This technique may be particularly useful in monitoring the recovery of a patient after angioplasty or placement of a stent. The blockage in an artery may return in a process called restenosis. Using a stethoscope to diagnose restenosis may allow for more frequent tests and the avoidance of multiple angiograms.

The preceding specific embodiments are illustrative of the practice of the invention. It is to be understood, therefore, that other expedients known to those skilled in the art or disclosed herein may be employed without departing from the invention or the scope of the claims.

Many embodiments of the invention have been described. Various modifications may be made without departing from the scope of the claims. These and other embodiments are within the scope of the following claims.

Claims

1. A method for detecting coronary artery disease using an electronic stethoscope, the method comprising:

recording acoustic data with an electronic stethoscope having an acoustic sensor placed over a fourth left intercostal space of a patient's chest;

applying one or more filters to the acoustic data and calculating a Fast Fourier Transform (FFT) of the data to produce FFT data;

analyzing the FFT data to detect a bell curve within a predefined frequency range indicative of coronary artery disease; and

generating an output when the bell curve is detected to indicate that the patient is likely to have coronary artery disease.

2. The method of claim 1, wherein analyzing the FFT data includes analyzing the FFT data with the electronic stethoscope to identify the bell curve.

3. The method of claim 1, wherein generating an output comprises updating a display of the electronic stethoscope.

4. The method of claim 1, further comprising communicating the acoustic data to an external computer via an interface module, and

wherein the steps of applying one or more filters and analyzing the FFT data are performed within the external computer.

5. The method of claim 1, wherein the acoustic sensor is a microphone.

6. The method of claim 1, wherein the acoustic sensor is a piezoelectric crystal.

7. The method of claim 1, wherein the acoustic data includes sounds generated during systole and diastole.

8. The method of claim 1, further comprising sending the acoustic data recorded with the electronic stethoscope to a general purpose computer for analyzes and outputting the indicator.

9. The method of claim 1, wherein one filter removes frequencies not related to stenosis from the acoustic data.

10. The method of claim 1, wherein one filter is a low pass filter that removes frequencies higher than 100 Hz.

11. The method of claim 1, further comprising displaying a graph of the processed data for manual diagnosis including plotting harmonic magnitudes on the y-axis and plotting harmonic frequencies on the x-axis.

12. The method of claim 11, wherein displaying the graph comprises:

displaying the graph in a first color, and

displaying the detected bell curve in a second color to aid recognition by a clinician.

13. The method of claim 1, wherein analyzing FFT data to identify a bell curve within a predefined frequency range includes identifying a bell curve between 50 and 80 Hz.

14. The method of claim 1, further comprising identifying the onset of the bell curve on or after 50 Hz.

15. The method of claim 1, further comprising identifying a downslope in the curve that ends at or before 80 Hz.

16. The method of claim 1, further comprising identifying a FFT peak in the bell curve greater than 2.5 units.

17. The method of claim 1, further comprising:

computing a total energy associated with the bell curve;

determining a percentage indicator of an extent of the artery stenosis based on the total energy; and

outputting the percentage indicator.

18. The method of claim 1, further comprising:

computing a total area under the bell curve;

determining a percentage indicator of an extent of the artery stenosis based on the total area; and

outputting the percentage indicator.

19. The method of claim 1, wherein the coronary artery disease is detected in the left anterior descending coronary artery.

20. An electronic stethoscope comprising:

a display;

an acoustic sensor;

one or more memories that stores acoustic data from the acoustic sensor and instructions for processing the acoustic data; and

a processor that applies one or more filters to the acoustic data and calculating a Fast Fourier Transform (FFT) of the data to produce FFT data,

wherein the processor analyzes the FFT data to detect a bell curve within a predefined frequency range, and generates an output on the display when the bell curve is detected to indicate that the patient is likely to have coronary artery disease.

21. The electronic stethoscope of claim 20, wherein the processor analyzes the FFT data to detect a bell curve having an onset on or after 50 Hz and a downslope at or before 80 Hz.

22. The electronic stethoscope of claim 20, wherein the processor analyzes the FFT data to identify a peak in the bell curve greater than 2.5 units.

23. The electronic stethoscope of claim 20, wherein the processor analyzes the FFT data to compute a total energy associated with the bell curve, determine a percentage indicator of an extent of the artery stenosis based on the total energy, and output the percentage indicator.

24. The electronic stethoscope of claim 20, wherein the processor analyzes the FFT data to computing a total area under the bell curve, determine a percentage indicator of an extent of the artery stenosis based on the total area; and output the percentage indicator.

25. A computer-readable medium comprising instructions to cause an electronic stethoscope having a processor to:

record acoustic data with an electronic stethoscope having an acoustic sensor placed over a fourth left intercostal space of a patient's chest;

apply one or more filters to the acoustic data and calculating a Fast Fourier Transform (FFT) of the data to produce FFT data;

analyze the FFT data to detect a bell curve within a predefined frequency range indicative of coronary artery disease; and

generate an output when the bell curve is detected to indicate that the patient is likely to have coronary artery disease.